Impacto

Helianthus maximiliani

This chapter is formatted for the journal “The American Midland Naturalist”

The citation for this chapter is: Tetreault, H.M., Kawakami, T., Levy, C., and M.C. Ungerer. 2016. Low temperature tolerance in the perennial sunflower Helianthus maximiliani. The American Midland Naturalist. 175(1):91-102.

Abstract

Species distributed across diverse climate and thermal conditions represent opportune systems for studying tolerance of low temperature stress. We examined variation in cold

acclimation capacity and freezing tolerance among three natural populations (Texas, Kansas, and Manitoba) of the perennial sunflower species Helianthus maximiliani, originally collected across a 2134 km latitudinal transect in central North America. Tolerance to low temperatures was evaluated through leaf electrolyte leakage assays that quantify loss of cellular electrolytes into an aqueous medium due to plasma membrane damage. Freezing tolerance was highest for plants from the northernmost latitude (Manitoba population) under both non cold- acclimated and cold- acclimated experimental conditions. Individuals from Kansas and Texas populations exhibited lower freezing tolerance compared to Manitoba but did not differ from one another. All

populations retain the ability to increase freezing tolerance through cold acclimation, and effects of cold acclimation actually trended greater in populations from warmer regions (Texas and

71

Kansas). Freezing tolerance of Manitoba X Texas F1 hybrids was statistically indistinguishable from plants from the Texas population, suggesting patterns of genetic dominance for alleles in Texas populations. Analysis of flowering specimens from herbaria records of corresponding regional locations indicates considerable variation in flowering phenology whereby flowering occurs progressively earlier with increasing latitude. This phenological variation may provide an additional mechanism of coping with low temperature stress through temporal avoidance.

KEYWORDS: abiotic stress; cold acclimation; electrolyte leakage; freezing tolerance; phenology; avoidance

Introduction

Freezing temperatures represent an important abiotic stress for plants and limit species distribution patterns and opportunities for dispersal and colonization (Woodward, 1987).

Geographical variation in low temperature extremes can drive local adaptation within species as well, especially for taxa distributed broadly (Green, 1969; Casler et al., 2004; Saenz-Romero and Tapia-Olivares, 2008; Zhen and Ungerer, 2008a; Lee et al., 2012). Because of their sessile lifestyle and inability to escape ambient climate conditions, plants provide a powerful

experimental system to examine strategies of coping with low temperature stress, with direct relevance to ecological and evolutionary population dynamics and agriculture.

For many temperate plant species, maximum freezing tolerance is enhanced via

acclimation to low but nonfreezing temperatures. This phenomenon, known as cold acclimation, is marked by major cellular biochemical changes enabling plants to withstand temperatures

72

several degrees colder than non cold- acclimated controls (Guy, 1990; Xin and Browse, 2000; Iba, 2002; Smallwood and Bowles, 2002; Guy et al., 2008). Cold- acclimation represents an inducible change and likely evolved in response to seasonal changes in which low, nonfreezing temperatures portend colder temperatures that are potentially more harmful. Genes and gene pathways regulating plant cold- acclimation have been identified in the model plant species Arabidopsis thaliana (Gilmour et al., 1998; Liu et al., 1998; Xin and Browse, 2000; Chinnusamy et al., 2003; Vogel et al., 2005; Agarwal et al., 2006; Van Buskirk and Thomashow, 2006; Doherty et al., 2009; Thomashow, 2010) with some genes shown to exhibit functional variability among natural A. thaliana accessions subjected to different historical selection pressures for freezing tolerance (Zhen and Ungerer, 2008b).

Avoidance mechanisms also represent feasible means through which organisms can cope with abiotic stress. Such mechanisms are relevant to plant species via seasonal growth patterns and reproductive timing events that minimize the probability of encountering stressful

environments and/or conditions (Heide, 1994; Bennington and Mcgraw, 1995; Griffith and Watson, 2005; Heschel and Riginos, 2005). Tolerance and avoidance mechanisms of abiotic stress need not be mutually exclusive and both may function in natural plant populations (Geber and Dawson, 1997; Heschel and Riginos, 2005).

Helianthus maximiliani is a diploid perennial sunflower species that flowers in late summer-fall (Heiser et al., 1969; Schilling, 2006). Though distributed widely in North America, populations are found in highest concentration in mid-continental regions between Texas, U.S.A. and Manitoba, CA (Schilling, 2006). Helianthus maximiliani exhibits steep latitudinal clines in

73

multiple morphological and life history traits that are hypothesized to be driven by variation in photoperiod and climate (Kawakami et al., 2011). A role for natural selection in shaping clinal variation in this species is supported by patterns of phenotypic differentiation among populations that exceed neutral expectations estimated from putatively neutral molecular markers, i.e., Qst — Fst analysis (Leinonen et al., 2013).

In the current report, we test whether H. maximiliani populations from three climatically diverse regions in central North America (Texas, Kansas and Manitoba) differ in their tolerance to freezing temperatures and whether effects of cold acclimation treatment differentially impacts freezing tolerance among these groups. We explore basic genetic aspects of cold acclimation in H. maximiliani by examining cold acclimation capacity and freezing tolerance of F1 hybrids derived from a cross between Manitoba and Texas plants. We demonstrate plants from the northernmost population (Manitoba) predictably display highest freezing tolerance under most treatment combinations but that cold acclimation effects persist (and are of greater magnitude under the experimental temperatures assayed) for plants from warmer regions (Texas and Kansas). Patterns of freezing tolerance of Manitoba x Texas F1 hybrids more closely resemble plants from Texas, suggesting dominance of alleles in Texas populations. We additionally examine patterns of H. maximiliani flowering phenology based on herbaria records in the native regions of these populations in light of corresponding seasonal temperature data. We reveal highly different phenology among populations that also may serve as an avoidance mechanism of low temperature stress.

74

Methods

Sunflower populations and growing conditions

Seeds from natural populations of H. maximiliani were collected in the field or obtained from the USDA National Plant Germplasm System (Table 4.1). Seeds obtained from the USDA are derived directly from wild- collected populations. Manitoba x Texas F1 hybrids were

generated with a Manitoba individual serving as maternal parent and assayed for freezing tolerance alongside individuals from the natural populations. All plants were grown in 8 in pots with a 1:1 mixture of Metro- mix 350: all- purpose sand under a 14:10 h light:dark cycle and ambient temperature in the Kansas State University greenhouses. Watering was conducted daily or as needed and fertilization with a weak nutrient solution (N:P:K = 15:30:15) was provided once per week. Plants were randomly positioned across three 1.2 x 2.4 m greenhouse benches. Plants receiving a cold- acclimation treatment prior to assays of freezing tolerance were placed in a 4 C walk- in chamber for 6 d, where they experienced constant low ambient light conditions and were watered as needed.

Electrolyte leakage assay

Damage to plant tissue from exposure to freezing stress can be quantified through loss of cellular electrolytes into an aqueous medium due to damage to cellular plasma membranes (Sukumaran and Weiser, 1972). Terminal 5 cm sections of leaves were excised and placed in individual 50 ml plastic Corning tubes containing 1 ml of ddH2O with the cut side of the leaf sample facing the tube bottom. All tubes were kept on ice during sample collection. Samples were subjected to freezing temperatures in an ESPEC ESU-3CA Platinous series environmental

75

test chamber (ESPEC North America, Hudsonville, Michigan, USA). Samples from non cold- acclimated plants were subjected to freezing stress at -4 C and -5 C; samples from cold- acclimated plants were subjected to freezing stress at -5 C and -6 C. These temperatures were selected based on preliminary experiments surveying tolerances across different temperature ranges. Assays of non cold- acclimated and cold- acclimated individuals at the same temperature (i.e., -5 C) enabled determination of the effects of cold- acclimation on freezing tolerance. To facilitate ice nucleation during periods of cooling, ice chips were added to Corning tubes when the chamber temperature reached -1 C. Minimum temperatures were maintained for 3 h and the duration of cooling (-1 C to experimental minimum temperature) and warming periods

(experimental minimum temperature to 4 C) was 2.5 h and 5 h, respectively. Rates of

temperature change thus ranged 1.2 – 2 C h-1 _{during cooling and 1.6 – 2 C h}-1 _{during warming.}

Following freezing stress treatments, samples were removed from the environmental chamber, ddH2O was added to fully cover the leaf samples, and tubes were placed on a platform shaker at 180 rpm for 24 h. The following day ion conductivity of the solution in each tube was measured using a Mettler Toledo FiveEasy conductivity meter (Mettler-Toledo, Columbus, Ohio, USA). All samples were measured twice and the mean of the two measurements used in subsequent analyses. Following initial measurements, samples were placed at -20 C for 24 hours to fully rupture cells and maximize freezing- based tissue damage. Samples were subsequently thawed and placed on a platform shaker at 180 rpm for 1 h. Ion conductivity was measured again by the same protocol, and the ratio of ion conductivities at the two temperatures (freezing stress assay temperature/ -20C) was used as a metric of relative electrolyte leakage for the sample. Lower and higher ratios of electrolyte leakage thus correspond to higher and lower tissue

76

freezing tolerance, respectively. For each acclimation temperature combination, 17 biological replicates of each genotype (i.e., population or F1 hybrid) were assayed in two replicate sets (n=9 and n=8, sequentially). For all assays, fully expanded and healthy leaves of similar age were harvested from the central region of the plant stem.

Electrolyte leakage data were analyzed by mixed model analysis of variance (ANOVA) using JMP version 5.0.1a (SAS Institute, Cary, North Carolina, USA). Cold- acclimated and non cold- acclimated plants were tested over different temperature ranges and analyzed separately

according to the model , where P is population, T is

temperature, and R is replicate. An additional model, ,

was evaluated examining cold- acclimated and non cold- acclimated plants assayed at -5 C, where terms are the same as those above and where A represents acclimation treatment. Square brackets represent nested terms. Replicate was treated as a random effect with all other main effects treated as fixed.

Natural phenological and climate data

Historical flowering time data for H. maximiliani in the relevant collection locations were obtained from herbarium specimens, which can serve as a proxy for seasonal flowering times and flowering durations for locally collected samples. Data for Texas plants were obtained from The University of Texas at Austin Plant Resources Center (TEX-LL) in electronic format (www.biosci.utexas.edu/prc/databases.html), limiting the query to individuals in anthesis. Data for Kansas and Manitoba plants were collected manually by recording collection dates of herbarium specimens in flower, from the Kansas State University Herbarium (KSC) and the

y = u+ P +T + P ´T + R[T]+ R´ P[T]

77

University of Manitoba Herbarium (WIN), respectively. Only specimens collected within the state or province boundaries of the relevant herbaria were utilized and only one specimen per collector per collection date was retained for analysis. All flowering dates were converted to Julian calendar days.

Seasonal first frost information for Manitoba is based on climate information reported in Environment Canada (http://www.ec.gc.ca/). Corresponding data for Manhattan, Kansas and Austin, Texas were obtained from NOAA Satellite and Information Service

(http://www.ncdc.noaa.gov/cdo-web/). These locations are centrally located relative to the collection locations of the specimens analyzed. Seasonal first frost data represent averages from 1951 to 1980, and define ‘light freeze’ events (29 to 32 F) with 50% possibility of frost occurring before or after.

Results

Freezing tolerance variation among populations

Populations of H. maximiliani from locations in Manitoba, Kansas and Texas experience appreciably different temperature conditions during the growing season and likely face different selection pressures for tolerance to low temperature. In assays of freezing-induced leaf

electrolyte leakage, ANOVA revealed significant effects of Population and Temperature in analyses of non cold- acclimated plants (Table 4.2) and significant effects of Population and Replicate for cold- acclimated plants (Table 4.3). None of the interaction terms in either

statistical model were significant. Significantly lower electrolyte leakage (higher tolerance) was observed for Manitoba plants versus other populations under non cold-acclimated conditions at

78

both -4 C and -5 C (Figure 4.1A). Under cold- acclimated conditions, significantly lower

electrolyte leakage was observed for Manitoba plants versus Texas plants at -6 C, but significant differences were not observed among populations at -5 C (Figure 4.1B). F1 hybrids derived from a Manitoba x Texas inter-population cross displayed electrolyte leakage scores more similar to, and statistically indistinguishable from Texas plants across all acclimation and temperature treatments (Figure 4.1A, B). For all populations and acclimation conditions, colder temperatures resulted in higher electrolyte leakage scores (Figure 4.1A, B).

Effects of cold-acclimation on freezing tolerance

Assays of cold- acclimated and non cold- acclimated plants subjected to freezing stress at -5 C enabled examination of acclimation effects on leaf freezing tolerance and whether

differences among populations exist with regard to this inducible response. Analysis of this subset of the data revealed significant effects of Population and Acclimation (Table 4.4), with all other effects nonsignificant. For all populations and the F1 hybrids, cold- acclimation resulted in lower leaf electrolyte leakage scores [higher tolerance] (Figure 4.2). Populations/F1 hybrids did not differ in the effects of cold- acclimation treatment on enhancing freezing tolerance (F = 2.1394; P = 0.1965 for the Population*Acclimation interaction; Table 4.4) although the difference in non cold- acclimated versus cold- acclimated leaf electrolyte leakage trended consistently higher for Kansas and Texas populations and F1 hybrids versus Manitoba (Figure 4.2). Indeed, differences were not observed among populations subjected to -5 C following cold acclimation (Figure 4.1B) whereas Kansas and Texas populations exhibited significantly higher electrolyte leakage values versus Manitoba at this same temperature under non cold acclimated conditions (Figure 4.1A).

79 Regional phenologies

Similar numbers of H. maximiliani records were analyzed from each of three herbaria: University of Manitoba-WIN (n=38), Kansas State University-KSC (n=35), and University of Texas-TEX-LL (n=28). Considerable variation in seasonal phenology was observed among locations, with earliest flowering dates recorded for Manitoba plants (mean = 220±4.9 Julian days, range = 182 to 252), followed by Kansas plants (mean = 251±5.1 Julian days, range = 195 to 299), and Texas plants (mean = 267±5.7 Julian days, range = 98 to 322). Flowering periods largely precede 30 y means of autumn first frost dates within but not across regions (e.g., Texas or Kansas phenology data compared with Manitoba climate data or Texas phenology data compared with Kansas climate data) [Figure 4.3].

Discussion

Population variation in freezing tolerance

Ambient temperatures encountered by natural plant populations vary predictably with latitude and selection pressures for tolerance to low temperature are expected to be stronger for populations from colder climates (Dionne et al., 2001; Shahba et al., 2003; Zhen and Ungerer, 2008a). We tested this prediction in the broadly distributed perennial sunflower species H. maximiliani and found that both with and without cold-acclimation pretreatments, plants from the highest latitude (Manitoba) exhibited the lowest levels of leaf electrolyte leakage (highest tolerance) compared to plants from Kansas and Texas populations, though not all comparisons were significant in post hoc tests.

80

No differences in leaf electrolyte leakage were observed between samples from Kansas and Texas populations for any of the experimental temperature/pretreatment combinations, suggesting that tolerance to low temperature for these populations is indistinguishable, at least under the conditions utilized in the current study. Lack of observed differences in this trait between plants from Kansas and Texas populations lies in contrast to large differences across the same transect for life history and size- related traits such as flowering time, dry biomass, and stem diameter (Kawakami et al., 2011). These results suggest that physiological, morphological, and life history characters may be subject to different selection pressures across the species range.

While plant freezing tolerance is a classic quantitative trait with a complex genetic basis (Thomashow, 1999), F1 hybrids derived from a cross between plants from Manitoba and Texas populations displayed levels of leaf electrolyte leakage more similar to and statistically

indistinguishable from Texas plants, suggesting alleles of large effect from Texas plants exhibit dominance with respect to this phenotype. Similar patterns have been observed in segregating F2 hybrids between Manitoba and Texas H. maximiliani plants of the same populations for traits associated with plant architecture and growth rate (Kawakami et al., 2011). It is currently

unknown whether these traits collectively have a shared genetic basis though genetic correlations for freezing tolerance and life history traits have been observed in other plant species (Agrawal et al., 2004).

81 Effects of cold-acclimation across populations

Cold- acclimation pretreatment of 4 C for 6 days resulted in reduced leaf electrolyte leakage (enhanced freezing tolerance) for all groups under study. Populations/F1 hybrids did not differ in the magnitude of this effect as determined by a nonsignificant Population*Acclimation interaction term (F = 2.1394, P = 0.1965; Table 4.4) although the difference in leaf electrolyte leakage under these different conditions trended lower for Manitoba plants versus others (Figure 4.2). This is an unexpected and interesting finding given that, in other plant species, individuals from colder environments have been shown to exhibit a stronger cold- acclimation response (Hannah et al., 2006). This result should be interpreted with caution, however, as the lesser ability of Manitoba plants to undergo cold- acclimation in the current study could be attributable to lower levels of leaf electrolyte leakage under non cold-acclimated conditions on account of higher intrinsic freezing tolerance for this population (Figure. 4.1A, 4. 2) and thus be an artifact of the particular assay conditions utilized. It is clear, however, that all assayed populations of H. maximiliani possess the relevant biochemical machinery to undergo this important physiological change. Populations from southern (warmer) climates have not lost this capacity, despite the potential for relaxed selection on cold-acclimation capacity in warmer climates (Zhen and Ungerer, 2008b).

Results presented here also contrast with previous reports describing an absence of cold acclimation capacity in domesticated varieties of the common sunflower Helianthus annuus (Hewezi et al., 2006; Allinne et al., 2009). It is currently unknown whether wild accessions of H. annuus also lack this capacity or if absence of this response in domesticated varieties could be a consequence of reduced genetic variability following domestication (Mandel et al., 2011). Given

82

similar widespread distributions of H. maximiliani and H. annuus and a relatively young age of the genus as a whole (Kane et al., 2013), the latter hypothesis seems more tenable, but further experiments are required to answer this question with certainty.

The genetic basis of cold- acclimation has not been investigated in H. maximiliani, although recent advances in next generation sequencing (NGS) have enabled cost- effective generation of genome-level resources for this nonmodel species. A recent analysis of

transcriptomes of Manitoba and Texas populations of H. maximiliani (Kawakami et al., 2014) revealed homologs of several Arabidopsis thaliana genes known to be involved in plant cold- acclimation, including important key regulators such as the C-repeat/dehydration responsive element binding factor 2 (CBF2) (Jaglo et al., 2001), calmodulin binding transcription activator 3 (CAMTA3) (Doherty et al., 2009), and inducer of CBF expression 1 (ICE1) (Chinnusamy et al., 2003). These and other identified homologs represent excellent ecological ‘candidate loci’ for future investigations of the molecular underpinnings of cold- acclimation and freezing tolerance variation among populations of H. maximiliani.

Phenological differences as a mechanism of abiotic stress avoidance

Natural history collections provide a useful resource for studying biological patterns and trends in nature (Lavoie and Lachance, 2006; Miller-Rushing et al., 2006; Robbirt et al., 2011; Panchen et al., 2012). Herbaria records of H. maximiliani collected from regional locations overlapping with the contemporary populations examined herein provide a rough approximation of historical flowering time information. Examined jointly with data on seasonal temperature change and averaged autumn first frost dates, it is possible to address whether regional

83

differences in flowering time may serve as an avoidance mechanism of low temperature stress. Earliest flowering in Manitoba populations and progressively later flowering in Kansas and Texas populations is consistent with differences in the length of the growing season in these locations and corroborates experimental flowering time data for this species under common garden conditions (Kawakami et al., 2011). For a given location, flowering records always precede historical autumn first frost dates (Figure 4.3). For comparisons across locations, however, this pattern is not upheld. For example, several Kansas specimens and the majority of Texas specimens were collected on dates following the average first frost for Manitoba

populations. Similarly, multiple Texas specimens were collected on dates following the average first frost for Kansas populations (Figure 4.3).

While these results are suggestive of reproductive timing events that avoid low temperature stress, such observational data are not without caveats. First, herbarium records,